Mustard Systems are a small statistical consultancy company of around 45 staff specialising in the modelling of sporting events. Applications of our research are widespread across multiple sectors including betting, insurance, finance, media as well as giving advice to coaches and owners of sporting teams. Our aim is to predict sport as accurately as possible, and we achieve this by combining large, rapidly changing datasets with home grown software and statistical models.
We enjoy a fast-paced, ever-changing environment, offering fresh challenges which are both rewarding and enjoyable. Our ambitions are high, aiming to create best-in-class systems in a highly competitive field.
In the coming year, we will be working on expanding our trading strategies and data feeds. This will include:
- Expanding the number of sports and markets we cover
- Scaling up our price feed infrastructure to cover many more sports, markets and customers
- Improving the traders’ visibility of our systems’ activity to provide more insight into their operation
You will help improve our systems, working directly with the users of your software. Those users will be a mixture of quantitative analysts and sports traders, all of whom are clued up and provide intelligent feedback.
You will be given the opportunity to work independently and to take on responsibility for our trading strategies and price feeds from inception to delivery.
The technology stack you will be working with currently consists of:
- Python, mainly 3.5+ for new projects and 2.7 in some legacy systems
- Go for infrastructure services
- Linux servers
- PostgreSQL for data storage
- ZeroMQ and RabbitMQ for backend communication
- jQuery, React and vue.js for the frontend
You will have the opportunity to explore your own ideas and suggest new tools. Should we parse input data using something from the standard library? Would we be better off using a third party tool like ANTLR or Ragel? Could we use regular expressions, or would that just give us two problems? Would a quick sed script get the job done? Is the best solution to store output data in a relational database? Would a NoSQL store be a better fit for this problem, or could we use something hierarchical like HDF5? The answer will vary depending on the problem and you’ll be expected to investigate the options to find a good solution.
We’re an open source friendly company, preferring it where possible. But we’re also pragmatic enough to know that a closed source alternative may sometimes be a better choice. We believe that tests are important, but that they should be used where appropriate and in moderation. Running code that gets 95% of the job done is often preferable to perfect code with 100% test coverage that’s delivered 3 months late.
We offer a laid back working environment, flexible working hours, adjustable height desks, free cereals, fresh fruit, tea, coffee and fruit juice, as well as excellent benefits.
You will need a minimum of 1 year of commercial experience. You don’t need to know all the specific technologies we have used, although experience with multiple programming languages is a plus.
A software development background is preferred; mathematical and numerical skills are a bonus. Do you have a natural distrust of all things Microsoft and Apple? If so, get in touch.
- Competitive salary
- Working for an ambitious, fast-moving company that thinks a little differently from the industry standard
- Private pension plan
- Private health insurance
- Life Insurance
- Subsidised gym membership
Interested? Send your CV and covering letter to firstname.lastname@example.org.